| description abstract | This paper proposes a novel neural-network-based (NN-based) adaptive composite terminal sliding mode (TSM) control scheme for a class of spacecraft systems with mismatched disturbances and input uncertainties. The key point of the proposed scheme is that the multiple uncertainties are considered and well-handled by the combined second-order disturbance observer (SODO) method and an NN. First, for the purpose of handling the mismatched disturbances, a SODO is applied to obtain the exact estimation of the mismatched disturbances. Second, for satisfying the output constraints and avoiding the computational complexity, the constrained output signals are transformed into unconstrained ones by error transformation. Moreover, to ensure the fast convergence characteristic of the considered system, a SODO-based composite nonsingular terminal sliding mode (SODO-NTSM) manifold is designed. Furthermore, the external disturbances, the input uncertainties, and the unknown nonlinearities are considered by several adaptive laws. As a result, by combining the SODO-NTSM manifold, the external disturbances adaptive compensation laws, the adaptive compensation laws of the input uncertainties, and the NN weights adaptive update laws, the novel new NN-based adaptive composite TSM control scheme is established for the spacecraft system. The simulation results show that the proposed scheme presents the satisfactory control performance. | |
